Download 02/03/2010

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
Statistics
One-two sided test, Parametric and non-parametric test
statistics: one group, two groups, and more than two
groups samples
February 3, 2010
Jobayer Hossain, Ph.D. & Tim Bunnell, Ph.D.
Nemours Bioinformatics Core Facility
Nemours Biomedical Research
Class Objectives -- You will Learn:
1.
The difference between one- and two-sided tests of
hypotheses and when each is appropriate.
2.
The difference between parametric and nonparametric tests
and when each is most appropriate.
3.
How to select some common parametric and non-parametric
tests for quantitative and categorical variables involving:
4.
–
One-group
–
Two groups,
–
More than two groups
How to do these tests with SPSS
Nemours Biomedical Research
One- and two-sided tests of hypotheses
•
One-sided
– One direction of effect (e.g. mean efficacy of treatment group
is greater than the mean efficacy of placebo group)
– Greater power to detect difference in expected direction
•
Two-sided
– Effect could be in either direction (e.g. mean efficacy of
treatment group is not equal to the mean efficacy of placebo
group)
– More conservative
Nemours Biomedical Research
One sided and two sided test of hypothesis
One group
Two groups
One sided
A single mean differs
from a known value in a
specific direction. e.g.
mean > 0 or median > 0
Two means differ from
one another in a specific
direction. e.g.,
mean2 < mean1
median2 < median1
Two sided
A single mean differs
from a known value in
either direction. e.g.,
mean ≠ 0
median ≠ 0
Two means are not
equal. E.g.,
mean1 ≠ mean2
median1 ≠ median2
Nemours Biomedical Research
Parametric & Nonparametric Tests
•
Parametric Test
– Make certain assumptions about population distribution or
parameter of the population from which the sample is taken
– E.g. A normal population distribution and equality of population
variances among all groups being compared
•
Non-parametric Test
– Distribution free methods which do not rely on assumptions that the
data are drawn from a given probability distribution.
•
If data failed to meet assumptions non-parametric tests are
preferred
Nemours Biomedical Research
Parametric & Nonparametric tests for
Quantitative Variables
Hypothesis Testing Procedure
One group sample
Two-group sample
More than two groups sample
One sample t-test
Independent
Not Independent
Analysis of variance
Sign test
Two-sample t-test
Paired t-test
Kruskal Wallis test
Mann-Whitney U test
Wilcoxon Signed Rank test
Nemours Biomedical Research
Single group sample - one sample t-test
(Parametric)
•
Test for value of a single mean
•
E.g., to see if mean SBP of all AIDHC employees is 120 mm Hg
•
Assumptions
– Parent population is normal
– Sample observations (subjects) are independent
Nemours Biomedical Research
Single group sample-one sample t-test
(Parametric)
• Formula
Consider a sample of size n from a normal population with mean µ
and variance σ2, then the following statistic is distributed as
Student’s t with (n-1) degrees of freedom.
x−µ
t=
s/ n
x is the sample mean and s is the sample standard deviation
Nemours Biomedical Research
One group sample - Sign Test (Nonparametric)
•
Use:
(1) Compares the median of a single group with a specified value (instead
of single sample t-test).
•
Hypothesis: H0:Median = c
Ha:Median ≠ c
•
Test Statistic:
We take the difference of observations from median (xi - c). The number of
positive or negative difference follows a Binomial distribution. For a large
sample size, this distribution follows normal distribution.
Nemours Biomedical Research
One group sample: SPSS demonstration
•
One-sample t-test (Parametric)
– Analyze->Compare Means->One-Sample T-test. Then select a test
variable and a test value (value for H0) from this window and click
ok.
•
One-sample sign-test (Parametric)
– Analyze->Non-parametric->Binomial Test. Then select a test
variable, a cut point, and a value for test proportion (H0) and then
click ok.
Nemours Biomedical Research
Two-group (independent) samples - two-sample
t-statistic (Parametric)
• Use
– Test for equality of two means
• Assumptions
– Parent populations are normal
– Sample observations (subjects) are independent.
Nemours Biomedical Research
Two-group (independent) samples - two-sample
t-statistic (Parametric)
•
Formula (two groups)
– Case 1: Equal Population Standard Deviations:
• The following statistic is distributed as t distribution with (n1+n2 -2) d.f.
t=
( x1 − x2 )
1 1
Sp
+
n1 n2
The pooled standard deviation,
(n1 − 1) S12 + (n2 − 1) S 22
Sp =
n1 + n2 − 2
n1 and n2 are the sample sizes and S1 and S2 are the sample standard
deviations of two groups.
Nemours Biomedical Research
Two-group (independent) samples - two-sample
t-statistic (Parametric)
• Formula (two groups)
– Case 2: Unequal population standard deviations
• The following statistic follows t distribution.
t=
( x1 − x2 ) − ( µ1 − µ 2 )
s12 s22
+
n1 n2
• The d.f. of this statistic is,
(s
)
2
/ n1 + s / n2
v= 2
( s1 / n1 ) 2 ( s22 / n2 ) 2
+
n1 − 1
n2 − 1
2
1
2
2
Nemours Biomedical Research
Two-group (independent) samples - two-sample
t-statistic (Parametric) : SPSS demonstration
•
Analyze->Compare Means-> Independent-Samples T test. Then
select one or more test variables and a grouping variable.
Nemours Biomedical Research
Two-group (independent) samples- Wilcoxon
Rank-Sum Test (Nonparametric)
•
Use: Compares medians of two independent groups.
•
Corresponds to t-test for two independent sample means
•
Test Statistic:
Consider two samples of sizes m and n. Suppose N=m+n. Compute
the rank of all N observations. Then, the statistic,
W m= Sum of the ranks of all observations of the first variable (size
m).
Nemours Biomedical Research
Two-group (independent) samples- Wilcoxon
Rank-Sum Test (Nonparametric): SPSS demo
•
Analyze-> Nonparametric tests->Two-Independent-Sample
tests. Then select one or more test variables, a group variable,
and the test type Mann-Whitney U.
Nemours Biomedical Research
Two-group (matched) samples - paired tstatistic (Parametric)
• Use: Compares equality of means of two matched or
paired samples (e.g. pretest versus posttest)
• Assumptions:
– Parent population is normal
– Sample observations (subjects) are independent
Nemours Biomedical Research
Two-group (matched) samples - paired tstatistic (Parametric)
• Formula
– The following statistic follows t distribution with n-1 d.f.
d
t=
sd / n
Where, d is the difference of two matched samples and Sd is the standard
deviation of the variable d.
Nemours Biomedical Research
Two-group (matched) samples - paired tstatistic (Parametric): SPSS demo
• Analyze -> Compare Means->Paired-Samples T test.
Then select two dependent variables for variable1
(e.g. PLUC_pre) and variable2 (PLUC_post) and
then click ok.
Nemours Biomedical Research
Two-group (matched) samples Wilcoxon
Signed-Rank Test (Nonparametric)
•
USE:
– Compares medians of two paired samples.
•
Test Statistic
– Obtain differences of two variables, Di = X1i - X2i
– Take absolute value of differences, Di
– Assign ranks to absolute values (lower to higher), Ri
– Sum up ranks for positive differences (T+) and negative
differences (T-)
• Test Statistic is smaller of T- or T+ (2-tailed)
Nemours Biomedical Research
Two-group (matched) samples Wilcoxon
Signed-Rank Test (Nonparametric): SPSS
demo
•
Analyze -> Nonparametric tests -> Two-Related-Samples Test. Then
select two dependent variables for variable1(e.g. PLUC_pre) and
variable2 (PLUC_post), select Test type ‘Wilcoxon’ and click ok.
Nemours Biomedical Research
More than two independent samples: F statistic
(Parametric)
• Use:
– Compare means of more than two groups
– Test the equality of two variances.
Nemours Biomedical Research
More than two independent samples: F statistic
(Parametric)
•
Let X and Y be two independent Chi-square variables with n1 and n2
d.f. respectively, then the following statistic follows a F distribution with
n1 and n2 d.f.
X / n1
Fn1 ,n2 =
•
Y / n2
Let, X and Y are two independent normal variables with sample sizes
n1 and n2. Then the following statistic follows a F distribution with n1
and n2 d.f.
2
Fn1 ,n2
sx
= 2
sy
Where, sx2 and sy2 are sample variances of X and Y.
Nemours Biomedical Research
More than two independent samples: F statistic
(Parametric)
•
Hypotheses:
H0: µ1= µ2=…. =µ n
Ha: µ1≠ µ2 ≠ ≠ …. ≠ µn
Comparison will be done using analysis of variance (ANOVA)
technique.
ANOVA uses F statistic for this comparison.
The ANOVA technique will be covered in another class session.
Nemours Biomedical Research
More than two independent samples: F statistic
(Parametric) : SPSS demo
•
Create a new variable: PLUC.chng = PLUC.post – PLUC.pre
Transform-> Compute variable, then in this window, type the
name of the new variable under the Target variable (e.g.
PLUC.chng) and the expression PLUC.post – PLUC.pre under
the numeric expression
•
One-way ANOVA: Analyze -> Compare means -> One-way
ANOVA Select the dependent variable (e.g. PLUC.chng), and
factor (e.g. grp).
Nemours Biomedical Research
More than two groups: Nonparametric KruskalWallis Test
•
Compares median of three or more groups or (means of ranks
of three or more groups)
•
Rank the data ignoring group membership
•
Perform the one way ANOVA of ranks instead of data itself
•
SPSS Demo: Analyze -> Nonparametric Tests ->K Independent
Samples, then select one or more test variables and a group
variable and click ok.
Nemours Biomedical Research
Test statistics: Categorical variable
• Parametric
– Proportion Tests
– Chi-square Tests
• Nonparametric
– McNemar Test
– Kendall’s Tau
• We will discuss only Proportion and Chisquare tests
Nemours Biomedical Research
Proportion Tests
• Use
– Test for equality of two Proportions
• E.g. proportions of subjects in two treatment groups who
benefited from treatment.
– Test for the value of a single proportion
• E.g., to test if the proportion of smokers in a population
is some specified value (less than 1)
Nemours Biomedical Research
Proportion Tests
• Formula
– One Group:
z=
pˆ − p0
p0 (1 − p0 )
n
– Two Groups:
z=
pˆ1 − pˆ 2
1 1
ˆp(1 − pˆ )( + )
n1 n2
x1 + x2
where pˆ =
.
n1 + n2
Nemours Biomedical Research
Proportion Test: SPSS demo
• Analyze-> Nonparametric Tests-> Binomial. Then select a
test variable (e.g.Ped) and a value for Test proportion and
then click ok.
Nemours Biomedical Research
Chi-square Test
• USE
– Testing the population variance σ2= σ02.
– Testing the goodness of fit.
– Testing the independence/ association of attributes
• Assumptions
– Sample observations should be independent.
– Cell frequencies should be >= 5.
– Total observed and expected frequencies are equal
Nemours Biomedical Research
Chi-square Test
•
Formula: If xi (i=1,2,…n) are independent and normally distributed
with mean µ and standard deviation σ2, then,
2
 xi − µ 
2
is
a
χ
distribution with n d.f.


∑
σ 
i =1 
n
•
If we don’t know µ, 2then we estimate it using a sample mean and
 xi − x 
2
is
a
χ
distribution with (n - 1) d.f.


∑
σ 
i =1 
n
then,
Nemours Biomedical Research
Chi-square Test
• For a contingency table, we use the following chi- square
test statistic,
2
(
−
)
O
E
i
χ2 = ∑ i
, distribute d as χ 2 with (n - 1) d.f.
Ei
i =1
n
Oi = Observed Frequency
Ei = Expected Frequency
Nemours Biomedical Research
Chi-square Test
Female
O(E)
11 (10)
Total
Group 1
Male
O(E)
9 (10)
Group 2
8 (10)
12 (10)
20
Group 3
11 (10)
9(10)
20
30
30
60
Nemours Biomedical Research
20
Chi-square Test– calculation of expected
frequency
•
To obtain the expected frequency for any cell, use:
•
Corresponding (row total X column total) / grand total
•
E.g: cell for group 1 and female, substituting: (30 X 20 / 60) =
10
Nemours Biomedical Research
Chi-square Test: SPSS demonstration
•
Analyze->Descriptive statistics -> Crosstabs -> Pick row and column
variables, select other options and click ok
Nemours Biomedical Research
Thank you
Nemours Biomedical Research